process control a practical approach

Process Control A Practical Approach Process Control: A Practical Approach Myke King © 2011 John Wiley & Sons Ltd. ISBN: 978-0-470-97587-9 Process Control A Practical Approach Myke King Whitehouse Consulting, Isle of Wight, UK This edition first published 2011 2011 Ó John Wiley & Sons Ltd Registered office John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com. The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Library of Congress Cataloging-in-Publication Data King, Michael, 1951- Process control : a practical approach / Michael King. p. cm. Includes bibliographical references and index. ISBN 978-0-470-97587-9 (cloth) 1. Chemical process control. I. Title. TP155.75.K56 2011 660’.2815–dc22 2010034824 A catalogue record for this book is available from the British Library. Print ISBN: 9780470975879 e-PDF ISBN: 9780470976555 o-Book ISBN: 9780470976562 e-Pub ISBN: 9780470976661 Set in 10/12 Times Roman by Thomson Digital, Noida, India Printed in Singapore by Fabulous Printers Pte Ltd. Contents Preface ix About the Author xv 1. Introduction 1 2. Process Dynamics 3 2.1 Definition 3 2.2 Cascade Control 9 2.3 Model Identification 11 2.4 Integrating Processes 20 2.5 Other Types of Process 22 2.6 Robustness 24 2.7 Laplace Transforms for Processes 27 References 28 3. PID Algorithm 29 3.1 Definitions 29 3.2 Proportional Action 30 3.3 Integral Action 33 3.4 Derivative Action 35 3.5 Versions of Control Algorithm 39 3.6 Interactive PID Controller 41 3.7 Proportional-on-PV Controller 43 3.8 Nonstandard Algorithms 50 3.9 Tuning 51 3.10 Ziegler-Nichols Tuning Method 52 3.11 Cohen-Coon Tuning Method 56 3.12 Tuning Based on Penalty Functions 57 3.13 Manipulated Variable Overshoot 60 3.14 Lambda Tuning Method 61 3.15 IMC Tuning Method 63 3.16 Choice of Tuning Method 65 3.17 Suggested Tuning Method for Self-Regulating Processes 66 3.18 Tuning for Load Changes 66 3.19 Tuning for Unconstrained MV Overshoot 71 3.20 PI Tuning Compared to PID Tuning 72 3.21 Tuning for Large Scan Int erval 74 3.22 Suggested Tuning Method for Integrating Processes 76 3.23 Implementation of Tuning 78 3.24 Loop Gain 79 3.25 Adaptive Tuning 79 3.26 Initialisation 80 3.27 Anti-Reset Windup 81 3.28 On-Off Control 81 3.29 Laplace Transforms for Controllers 83 3.30 Direct Synthesis 85 References 88 4. Level Control 91 4.1 Use of Cascade Control 91 4.2 Parameters Required for Tuning Calculations 93 4.3 Tight Level Control 97 4.4 Averaging Level Control 100 4.5 Error-Squared Controller 105 4.6 Gap Controller 108 4.7 Impact of Noise on Averaging Contr ol 111 4.8 General Approach to Tuning 113 4.9 Three-Element Level Control 114 5. Signal Conditioning 117 5.1 Instrument Linearisation 117 5.2 Process Linearisation 119 5.3 Constraint Conditioning 122 5.4 Pressure Compensation of Distillation Tray Temperature 124 5.5 Pressure Compensation of Gas Flow Measurement 125 5.6 Filtering 126 5.7 Exponential Filter 127 5.8 Higher Order Filters 129 5.9 Nonlinear Exponential Filter 130 5.10 Averaging Filter 131 5.11 Least Squares Filter 132 5.12 Control Valve Characterisation 136 5.13 Equal Percentage Valve 137 5.14 Split-Range Valves 140 6. Feedforward Control 147 6.1 Ratio Algorithm 147 6.2 Bias Algorithm 151 6.3 Deadtime and Lead-Lag Algorithms 152 6.4 Tuning 155 6.5 Laplace Derivation of Dynamic Compensation 161 7. Deadtime Compensation 163 7.1 Smith Predictor 163 vi Contents 7.2 Internal Model Control 166 7.3 Dahlin Algorithm 167 References 168 8. Multivariable Control 169 8.1 Constraint Control 169 8.2 SISO Constraint Control 170 8.3 Signal Selectors 171 8.4 Relative Gain Analysis 174 8.5 Steady State Decoupling 177 8.6 Dynamic Decoupling 180 8.7 MVC Principles 184 8.8 Parallel Coordinates 187 8.9 Enhanced Operator Displays 188 8.10 MVC Performance Monitoring 189 References 195 9. Inferentials and Analysers 197 9.1 Inferential Properties 197 9.2 Assessing Accuracy 203 9.3 Laboratory Update of Inferential 208 9.4 Analyser Update of Inferential 210 9.5 Monitoring On-stream Analysers 212 Reference 214 10. Combustion Control 215 10.1 Fuel Gas Flow Correction 215 10.2 Measuring NHV 220 10.3 Dual Firing 222 10.4 Inlet Temperature Feedforward 223 10.5 Fuel Pressure Control 225 10.6 Combustion Air Control 227 10.7 Boiler Control 236 10.8 Fired Heater Pass Balancing 237 11. Compressor Control 243 11.1 Polytropic Head 243 11.2 Flow Control (Turbo-Machines) 246 11.3 Flow Control (Reciprocating Machines) 251 11.4 Anti-Surge Control 252 12. Distillation Control 259 12.1 Key Components 262 12.2 Relative Volatility 263 12.3 McCabe-Thiele Diagram 266 12.4 Cut and Separation 271 Contents vii 12.5 Effect of Process Design 281 12.6 Basic Controls 285 12.7 Pressure Control 285 12.8 Level Control 299 12.9 Tray Temperature Control 315 12.10 Pressure Compensated Temperature 325 12.11 Inferentials 335 12.12 First-Principle Inferentials 342 12.13 Feedforward on Feed Rate 344 12.14 Feed Composition Feedforward 348 12.15 Feed Enthalpy Feedforward 349 12.16 Decoupling 350 12.17 Multivariable Control 352 12.18 On-stream Analysers 360 12.19 Towers with Sidestreams 361 12.20 Column Optimisation 364 12.21 Optimisation of Column Pressure 366 12.22 Energy/Yield Optimisation 368 References 370 13. APC Project Executi on 371 13.1 Benefits Study 371 13.2 Benefit Estimation for Improved Regulatory Control 373 13.3 Benefits of Closed-Loop Real-Time Optimisation 380 13.4 Basic Controls 382 13.5 Inferentials 384 13.6 Organisation 385 13.7 Vendor Selection 389 13.8 Safety in APC Design 391 13.9 Alarms 392 References 393 Index 395 viii Contents Preface So why write yet another book on process control? There are already many published, but they are largely written by academics and intended mainly to support courses taught at universities. Excellent as some of these books are in meeting that aim, the content of many academic courses has only limited relevance to control design in the process industry. There are a few books that take a more practical approach but these usually provide only an introduction to the technologies. They contain enough detail if used as part of a wider engineering course but not enough for the prac titioner. This book aims more to meet the needs of industry. Most engineers responsible for the design and maintenance of control applications find daunting much of the theoretical mathematics that is common in the academic world. In this book we have aimed to keep the mathematics to a minimum. For example, Laplace transforms are only included so that the reader may relate what is in this book to what will be found in most theo retical texts and in the documentation provided by many DCS (distributed control system) vendors. They are not used in any of the control design techniques. And while we present the mathematical derivation of these techniques, to show that they have a sound engineering basis, the reader can skip these if too daunting and simply apply the end resu lt. The book aims to present techniques that have an immediate practical application. In addition to the design methods it describes any shortcuts that can be taken and how to avoid common pitfalls. The methods have been applied on many processes on a wide range of controllers. They should work! In addition to providing effective design methods, this book should improve the working practices of many control engineers. For example, the majority still prefer to tune PID (proportional, integral, derivative) controllers by trial and error. This is time-consuming and rarely leads to controllers performing as well as they should. This might be because of a justified mistrust of published tuning methods. Most do have serious limitations. This book addresses this and offers a method proven to be effective in terms of both controller performance and engineering effort. DCS include a wide array of control algorithms with many additional engineer-definable parameters. The DCS vendors are poor at explaining the purpose of these algorithms with the result that the industry is rife with misinterpretation of their advantages and disadvantages. These algorithms were included in the original system specification by engineers who knew their value, but this knowledge has not passed to the industry. The result is that there are substantial improvements that can be made on almost every process unit, surpassing what the control engineer is even aware of – let alone knows how to implement. This book addresses all the common enhancements. This book takes a back-to-basics approach. The use of MVC (multivariable controllers) is widespread in industry. Control engineering staff and their contractors have invested thousands of man-hours in the necessary plant testing and commissioning. Improving the basic controls is not usually an option once the MVC is in place. Improvements are likely to change the process dynamics and would thus involve substantial re-engineering of the MVC. Thus poor basic control remains the status quo and becomes the accepted standard to the point where it is not addressed even when the opportunity presents itsel f. This book raises the standard of what might be expected from the performance of basic controls. Before MVC, ARC (advanced regulatory control) was commonplace. MVC has rightly replaced many of the more complex ARC techniques, but it has been used by too many as the panacea to any control problem. There remain many applications where ARC out- performs MVC; but appreciation of its advantages is now hard to find in industry. The expertise to apply it is even rarer. This book aims to get the engineer to reconsider where ARC should be applied and to help develop the necessary implementation skills. However due credit must be given to MVC as a major step forward in the development of APC (advanced process control) techniques. This book focuses on how to get the best out of its application, rather than replicate the technical details that appear in many text books, papers and product documentation. The layout of the book has been designed so that the reader can progress from relatively straightforward conce pts through to more complex techniques appl ied to more complex processes. It is assumed that the new reader is comfortable with mathematics up to a little beyond high school level. As the techniques become more specific some basic knowledge of the process is assumed , but introductory information is included – particularly where it is important to control design. Heavily mathematical material, daunting to novices and not essential to successful implementation, has been relegated to the end of each chapter. SI units have been mainly used throughout but, where important and practical, conversion to imperial units is given in the text. Methods published in non-SI units have been included without change if doing so would make them too complex. The book is targeted primarily for use in the continuous process industry, but even predominantly batch plants have continuous controllers and often have sections of the process which are continuous. My experience is mainly in the oil and petrochemicals industries and, despite every effort being taken to make the process examples as generic as possible, it is inevitable that this will show through. However this should not be seen as a reason for not applying the techniques in other industries. Many started there and have been applied by others to a wide range of processes. It is hoped that the academic world will take note of the content. While some institutions have tried to make their courses more relevant to the process industry, practiti oners still perceive a huge gulf between theory and practice. Of course there is a place for the theory. Many of the modern control technologies now applied in the process industry are developed from it. And there are other industries, such as aerospace, where it is essential. The debate is what should be taught as part of chemical engineering. Very few chemical engineers benefit from the theory currently included. Indeed the risk is that many potentially excellent control engineers do not enter the profession because of the poor image that theoretical courses create. Further, those that do follow a career in process control, can find themselves working in an organisation managed by a chemical engineer- ing graduate who has no appreciation of what process control technology can do and its importance to the business. x Preface It is the nature of almost any engineering subject that the real gems of useful information get buried in amongst the background detail. Listed here are the main items worthy of special attention by the engineer because of the impact they can have on the effectiveness of control design. . Understanding the process dynamic s is essential to the success of almost every process control technique. These days there is very little excuse for not obtaining these by plant testing or from historically collected data. There are a wide range of model identification products available plus enough information is given in Chapte r 2 for a competent engineer to develop a simple spreadsheet-based application. . Often overlooked is the impact that apparently unrelated controllers can have on process dynamics. Their tuning and whether they are in service or not, will affect the result of steptests and hence the design of the controller. Any changes made later can then severely disrupt controller performance. How to identify such controllers, and how to handle their effect, is described in Chapters 2 and 8. . Modern DCS include a number of versions of the PID controller. Of particular importance in the proportional-on-PV algorithm. It is probably the most misunderstood option and is frequently dismissed as too slow compared to the more conventional proportional-on-error version. In fact, if properly tuned, it can make a substantial improvement to the way that process disturbances are dealt with – often shortening threefold the time it takes the process to recover. This is fully explained in Chapter 3. . Controller tuning by trial and error should be seen as an admission of failure to follow proper design procedures, rather than the first choice of technique. To be fair to the engineer, every published tuning technique and most proprietary packages have serious limitations. Chapter 3 presents a new technique that is well proven in industry and gives sufficient information for the engineer to extend it as required to accommodate special circumstances. . Derivative action is too often excluded from controllers. Understandably introducing a third parameter to tune by trial and error might seem an unnecessary addition to workload. It also has a poor reputation in the way that it amplifies measurement noise, but, engineered using the methods in Chapter 3, it has the potential to substantially lessen the impact of process disturbances. . Tuning level controllers to exploit surge capacity in the process can dramatically improve the stability of the process. However the ability to achieve this is often restricted by poor instrument design, and, often it is not implemented because of difficulty in convincing the plant operator that the level should be allowed to deviate from SP (set-point) for long periods. Chapter 4 describes the important aspects in sizing and locating the level transmitter and how the conventional linear PID algorithm can be tuned – without the need even to perform any plant testing. It also shows how nonlinear algorithms, particularly gap control, can be set up to handle the situation where the size of the flow disturbances can vary greatly. Preface xi [...]... typically captures the large majority of the available process control benefits The main technology applied here is the multivariable controller (MVC) Because of its relative ease of use and its potential impact on profitability it has become the focus of what is generally known as advanced process control (APC) In fact, as a result, basic control and ARC have become somewhat neglected Many sites (and many... unstable can also be applied to controllers that have saturated This means that the controller output has reached either its minimum or maximum output but not eliminated the deviation between PV and SP It can also be applied to a controller using a discontinuous on-stream analyser that fails Such analysers continue to transmit the last measurement until a new one is obtained If, as a result of analyser failure,... The data collection interval can be quite long We will show later that steady state is virtually reached within y þ 5t Assuming we need around 30 points to achieve a reasonably accurate fit and that we make both an increase and a decrease in the MV, then collecting data at a one-minute interval would be adequate for a process which has time constants of around two or three minutes This model identification... of samples are collected to obtain dynamic behaviour, for example if an onstream analyser is temporarily out of service or its installation delayed The additional laboratory testing generated may be substantial compared to the normal workload If the laboratory is not expecting this, then analysis may be delayed for several days with the risk that the samples may degrade The most accurate way of determining... the process If the PV is subject to noise, small disturbances will be difficult to analyse accurately The change in PV needs to be at least five times larger than the noise amplitude This may cause an unacceptable process disturbance Dynamics, as we shall see later in Chapter 6, are not only required for changes in the MV but also for disturbance variables (DV ) It may be that these cannot be changed as... that, for large values of n, the response becomes closer to a step change This confirms that a series of lags can be approximated by deadtime But it also means that deadtime can be approximated by a large number of small lags We will cover, in Chapters 6, 7 and 8, control schemes that require a deadtime algorithm If this is not available in the DCS then this approximation would be useful 2.4 Integrating... simple fired heater as shown in Figure 2.1 It has no automatic controls in place and the minimum of instrumentation – a temperature indicator (TI) and a fuel control valve The aim is to ultimately commission a temperature controller which will use the temperature as its process variable (PV ) and the fuel valve position as it manipulated variable (MV ) Figure 2.2 shows the effect of manually increasing the... between the control engineer and others working on the instrumentation and system The simplistic approach is to assign all hardware to these staff and all configuration work to the control engineer But areas such as algorithm selection and controller tuning need a more flexible approach Many basic controllers, providing the tuning is reasonable, do not justify particular attention Work on those that do requires... demonstrates that automatically updating the inferential bias with laboratory results will generally aggravate the problem Simple monitoring of on-stream analysers, described in Chapter 9, ensures that measurement failure does not disrupt the process and that the associated reporting tools can do much to improve their reliability and use Preface xiii Compensating fuel gas flow measurement for variations... to better handle process disturbances Chapter 10 shows how the benefits of both approaches can be captured Fired heater pass balancing is often installed to equalise pass temperatures in order to improve efficiency Chapter 10 shows that the fuel saving is negligible and that, in some cases, the balancing may accelerate coking However there may be much larger benefits available from the potential to debottleneck . Process Control A Practical Approach Process Control: A Practical Approach Myke King © 2011 John Wiley & Sons Ltd. ISBN: 978-0-470-97587-9 Process Control A Practical Approach Myke. obtaining the process dynamics for MVC packages, most will take a much less analytical approach to regulatory controls. This chapter aims to demonstrate that process dynamics can be identified easily. that are used to assess their accuracy are flawed and can lead the engineer into believing that their performance is adequate. It also demonstrates that automatically updating the inferential